Abstract
Background
The association between headache and paranasal sinus disease is still unclear. Because of symptom overlap, the two conditions are not easily studied on the basis of symptoms alone. The aim of the present study was to investigate whether paranasal sinus opacification on magnetic resonance imaging (MRI) was associated with migraine, tension-type headache (TTH) or unclassified headache.
Methods
This was a cross-sectional study of 844 randomly selected participants (442 women, age range 50–65 years, mean age 57.7 years). Based on 14 headache questions, participants were allocated to four mutually exclusive groups: migraine, TTH, unclassified headache or headache free. On MRI, opacifications as mucosal thickening, polyps/retention cysts and fluid in the five paired sinuses were measured and recorded if ≥1 mm. For each participant, opacification thickness was summed for each sinus and, in addition, a total sum of all sinuses was calculated. Opacification in each sinus was compared between headache-free participants and the headache groups using non-parametric tests, and the total sum was compared by logistical regression.
Results
No significant association was found between paranasal sinus opacification and headache in general, nor when headache was differentiated into migraine, TTH and unclassified headache. This was also true in separate analyses of mucosal thickening and fluid and of opacification from each paranasal sinus.
Conclusion
Migraine, TTH and unclassified headache were found not to be associated with an increased degree of paranasal sinus opacification at MRI.
Keywords
Introduction
Headache is a common complaint in the general population (1) and represents a large socioeconomic burden for society (2). The term ‘sinus headache’ is commonly used when paranasal sinus disease is thought to cause headache (3). Specialists consider this a rare cause of headache, but nevertheless, a large proportion of patients diagnose themselves with ‘sinus headache’ (4), and most of these patients suffer from migraine (5).
It has been shown that patients with chronic rhinosinusitis have a significantly higher risk of chronic headache compared to the general population (6). At the same time, headache sufferers without rhinosinusitis (negative computed tomography or nasal endoscopy) have an increased level of symptoms on the 22-item sino-nasal outcome test (SNOT 22) compared to controls (7), and some symptoms of sinus disease, such as a stuffy nose, rhinorrhoea and facial pain, can also occur in migraine (8). Thus, because of this symptom overlap, it is problematic to study the association between sino-nasal disease and headache solely on the basis of symptoms. Acute and chronic rhinosinusitis are accepted as causes of headache in the International Classification of Headache Disorders, 3rd edition beta (9) if headache follows the clinical course of rhinosinusitis, as verified by endoscopy, imaging or clinical examination.
Very few studies have used diagnostic imaging of the paranasal sinuses in large populations in order to evaluate the association between sino-nasal disease and headache. The few that exist have either included a selected group of patients (10,11), have not included a control group (12) or have not differentiated between migraine and tension-type headache (TTH) (11,13).
The aim of this cross-sectional study was to investigate the association between paranasal sinus opacification on magnetic resonance imaging (MRI) and headache, including migraine, TTH and unclassified headache in the general population, with participants recruited from a large epidemiological study in Norway (the HUNT study-MRI).
Material and methods
Study sample
The Nord-Trøndelag health study (the HUNT study) is a large-scale epidemiological study that was conducted in three main waves: HUNT1 (1984–1986), HUNT2 (1995–1997) and HUNT3 (2006–2008) (14). In each wave, the entire population of the Norwegian county of Nord-Trøndelag aged 20 years or older was invited to participate. In total, 77,212 (89%), 65,237 (70%) and 50,807 (54%) individuals participated in HUNT1, HUNT2 and HUNT3, respectively. Detailed information about the participants' health status was collected through questionnaires, biomedical measurements and blood samples (14).
A total of 1560 participants were randomly selected for a MRI of the head as part of the HUNT-MRI study. Criteria were age 50–65 years at the time of the HUNT3 screening, living fewer than 45 minutes of travel from the MRI examination centre in Levanger and participation in all three HUNT studies (n = 14,033). Of these, 554 did not respond to the invitation or were excluded owing to general MRI contraindications, stratification or because the examination had to be terminated during the scan, leaving 1006 participants who gave written consent and had the MRI. Details of the HUNT-MRI study are described elsewhere (15–17). After the MRI examinations were performed in the period July 2007 to December 2009, 24 participants were excluded from this particular study because of low image quality of the paranasal sinuses. Of the 982 participants where image quality was sufficient (18), 138 participants were excluded due to missing information on headache. Finally, 844 participants were included (54% participation rate): 442 women and 402 men, age range 50–66 years, mean age 57.7 years.
Headache diagnosis
The questionnaire in HUNT3 consisted of 14 questions regarding headache. All participants answering ‘yes’ to the screening question, ‘Have you suffered from headache during the last 12 months?’ were asked to fill out 13 subsequent questions, which were designed to determine whether the person had migraine and TTH. The questions reflected the International Classification of Headache Disorders, 2nd edition (ICHD-II) criteria for these disorders (19), with three exceptions: migraine was also accepted when the attacks lasted less than four hours, as the HUNT study did not specifically ask about untreated attacks. Sensitivity to light and sound was covered in one question, and only participants experiencing headache on ≥1 day per month were categorised as having TTH, as the specificity was higher in frequent and chronic TTH. The diagnoses were mutually exclusive. Unclassified headache was an exclusion diagnosis defining those who gave a positive answer to the screening question for headache, but lacking the necessary characteristics to be classified as having migraine or TTH. Individuals who answered ‘no’ to the screening question were classified as ‘headache free’ and regarded as the reference group. The validity of these questionnaire-based diagnoses has been reported previously (20): for any headache, Hagen et al. found the sensitivity to be 88% and the specificity to be 86% (kappa value: 0.70, 95% confidence interval (CI): 0.61–0.79); for migraine, the sensitivity was 51% and the specificity was 95% (kappa value: 0.50, 95% CI: 0.32–0.68); and for TTH ≥ 1 day per month, the sensitivity was 96% and the specificity was 69% (kappa value: 0.44, 95% CI: 0.30–0.58).
MRI
MRI was performed using a 1.5-Tesla HDx scanner (Sigma, GE Healthcare, Waukesha, WI) equipped with an eight-channel head coil and software version pre-14.0M4. The scan protocol included axial T2-weighted (w) images, T1w magnetisation-prepared rapid acquisition gradient echo volume, scan axial T2w, T2* and fluid-attenuated inversion recovery sequences and a time-of-flight 3D angio-sequence. In this study, measurements were performed using the axial T2w images and coronal, sagittal and axial T1w images. The acquisition parameters of these sequences have been published previously (18).
MRI readings
MRI readings were obtained in the period April 2012 to July 2013 using a DICOM reader (Osirix version 3.2.4, 32 bit; Osirix Foundation, Geneva, Switzerland). Two physicians performed MRI readings independently, one specialised in paranasal sinus radiology (HBE) and one a resident in an ear, nose and throat department with 4 years of experience (AGH). Both were blinded to all participant data. Discrepancy in measurements or interpretation occurred in 21% of the cases. In these cases, the MRIs were re-examined by both examiners and a consensus was reached.
Opacification in the five paired paranasal sinuses (maxillary, anterior and posterior ethmoid, frontal and sphenoid on the left and right side) were measured at their maximum thicknesses in millimetres and recorded if ≥1 mm, as opacification of <1 mm was considered normal or insignificant.
The three categories of opacifications were mucosal thickening, polyps/retention cysts and fluid. Polyps and retention cysts were merged as one category, as they cannot be unambiguously differentiated on MRI (21). Details of how these categories of opacifications were defined have been provided previously (18). If more than one category of opacification occurred in the same sinus, each was measured. For each participant, the thickness of each category was measured for each paranasal sinus. These measurements were added, resulting in one sum for mucosal thickening, one sum for polyps/retention cysts and one sum for fluid for each participant. In addition, these sums were added to a total sum of opacification. Opacifications in each of the five paired paranasal sinuses (left and right side added) were also calculated.
Statistical analyses
With regard to the sum of fluid, the sum of mucosal thickening, the total sum of opacification and the thickness of opacifications in each paired sinus, the group of headache-free participants was compared with each of the headache groups (any headache, migraine, TTH and unclassified headache) using a non-parametric test (Mann–Whitney). This method was also used in sub-analyses for fluid in the maxillary and sphenoid sinuses (right and left side added). Fluid in the other paranasal sinuses was too rare to allow for statistical analyses. The outcomes ‘any headache’, ‘migraine’, ‘TTH’ and ‘unclassified headache’ versus ‘headache free’ were further investigated using logistic regression analysis, and the odds ratios (OR) with 95% CIs were estimated for the outcome by the total sum of paranasal sinus opacification.
As minor incidental findings are common on MRI of the paranasal sinuses and there is no consensus in the literature on how to define a pathological degree of opacification on MRI, percentiles were used to categorise the total sum of opacifications in the logistic regression analysis: minor degree of opacification (≤65th percentile); moderate degree of opacification (66th–85th percentile); and high degree of opacification (>85th percentile). For the total sum of opacification, this corresponded to cut-off values of 8 mm for the 66th percentile and 17 mm for the 86th percentile. Among the wide range of health-related information in the HUNT3 study, we have previously identified several important factors as being associated with headache (22) and paranasal sinus opacification (18). Potential confounders for this study were identified with the use of direct acyclic graphs (23), and in the adjusted analysis, the following variables were included as confounders: sex, age (5-year categories) and smoking (smoker/non-smoker).
The statistical analyses were conducted in PAWS Statistics, version 21 for Macintosh (SPSS Inc., Chicago, IL). p-values ≤ 0.05 were regarded as statistically significant.
The Regional Ethics Committee in central Norway approved the study (2011/2199-1).
Results
Of the 844 participants included in this study, 302 (36%) reported having suffered from headache during the past 12 months. The proportions of participants with migraine, TTH and unclassified headache are shown as a flowchart in Figure 1, and information regarding the confounders in provided in Table 1. The migraine group had a higher proportion of females (73%) and more smokers than the other headache groups.
Flowchart of participants in this study. Characteristics of the study population (n = 844) grouped into migraine, tension-type headache (TTH) and unclassified headache according to the International Classification of Headache Disorders, 2nd edition. Some participants are missing due to missing covariate information.
Mean thickness for the total sum of all opacification, the sum of mucosal thickening, the sum of fluid and the sum of opacification for each paired paranasal sinus (left and right) in millimetres (mean ± SD).
Compared to the headache-free group, the total sum of all opacification, the sum of mucosal thickening, the sum of fluid and the sum of opacification from each paired paranasal sinus (left and right side) were not significantly different in any of the headache groups (p > 0.12, Mann–Whitney test).
TTH: tension-type headache.
The association between paranasal sinus opacification on magnetic resonance imaging and headache (odds ratios with 95% confidence intervals) in unadjusted and adjusted analyses.
Adjusted for sex, age and smoking. Some participants are missing due to missing covariate information.
TTH: tension-type headache.
Discussion
To our knowledge, this is the first population-based study to investigate the association between headache as classified according to the ICHD-II criteria and opacification of the paranasal sinuses on MRI. Neither the opacification of a particular sinus nor a moderate or high degree of the total sum of opacifications were associated with headache in general, or migraine, TTH and unclassified headache. The implication of the study is that those who are headache sufferers in the general population are not more likely to have paranasal sinus opacifications than those without headache. It is important to emphasise that the study has little relevance to the acute situation (e.g. when a patient presents with symptoms of acute sinusitis) (24), since there were no data on symptoms – neither headache nor sino-nasal – on the day of the MRI scan.
The major strengths of this study were its population-based design, the large sample size and the use of validated headache diagnoses with good sensitivity and high specificity for both migraine and TTH (20). In the multivariate analyses, we were able to adjust for potential confounding factors. Furthermore, the investigated population was not considerably different from the general HUNT population: while the participants were non-selected in terms of health, they had somewhat higher levels of education and were less likely to be overweight or to have hypertension (15). It is also a strength that the MRI readings were performed independently by two readers who were blinded to all participant data.
However, some limitations should also be considered: the age range in the investigated groups was relatively restricted, so one should be cautious about generalisation of the results to other age groups; secondly, 138/982 (14%) of the MRI participants did not give information on headache and were excluded from the study. This is a potential source of selection bias, but we believe it is most likely non-differential, as headache was only one of several health-related topics in the HUNT questionnaire, which should reduce the likelihood of selective participation based on particular interest in headache.
A further limitation of the present study is that the MRIs were carried out 1–2 years after the information on headache was collected; thus, a change in headache characteristics during this time period could have led to misclassification. However, in another HUNT study (25), good agreement was found for the diagnosis of migraine and the status of being a headache sufferer, despite the fact that the questionnaire was filled in 5–9 months prior to the interview. Boardman et al. (26) found that the prevalence of headache was quite stable over time, but that some headache characteristics changed. Thus, it is likely that most participants with headache would have continued to report headache at the time of the MRI, but could have been classified into other headache groups. As we analysed the whole group of participants with any headache, and subsequently the subgroups of headache, we believe that we have taken this problem into account. We also recognise that some participants could have developed headache in the period between the headache report and the MRI. However, Lyngberg et al. (27) found a relatively low incidence of headache in the age group that was investigated in this study, leading us to believe that this is unlikely to have influenced our results.
There is no general consensus regarding the cut-off values for a pathological degree of opacification at MRI of the paranasal sinuses (28–30). Incidental findings have been shown to be quite common in a large proportion of the population (29). In the logistic regression analysis in this study, the cut-off value for a moderate degree of opacification was ≥8 mm. With this cut-off value, we believe that we avoid the problem of over-interpreting small findings in the paranasal sinuses. Since the cut-off value could be somewhat arbitrary and present a potential methodological bias, we measured all opacifications of ≥1 mm, as opacifications of <1 mm are likely to be normal. In this way, we were able to analyse opacification thickness as a continuous variable with the use of non-parametric tests in order to detect potential differences that would not be captured by using cut-off values in the regression analyses. An alternative way to classify the opacification on MRI would have been to use a scoring system, such as the Lund–Mackay scoring system (31). However, this scoring system is not validated for MRI and could have introduced new uncertainties. Other possible ways to evaluate the paranasal sinuses on imaging could have been to use a volumetric analysis of opacification, as this has been shown to have a higher correlation with sino-nasal symptoms (32), nasal endoscopy and quality of life (33) than the Lund–Mackay scoring system on computed tomography (CT). This method is novel and was not available at the time of MRI investigation in this study, but we believe that we have described opacification more precisely than with the use of scoring systems by measurements of opacification in millimetres. One further limitation was that we did not differentiate between different locations of opacification within the paranasal sinuses (e.g. those in the basis and those close to the upper jaw teeth versus those that could obstruct the sinus ostium). Such differentiation may be of clinical importance, and this distinction should be made in future studies on the relationship between paranasal sinus opacifications and headache.
To our knowledge, no population-based study has used paranasal sinus imaging to investigate the association between headache and sino-nasal disease. Our results are in concordance with studies in sinusitis patients, showing no association between headache/facial pain and opacification of the paranasal sinuses on CT (10,13). In these studies, headache was not differentiated into migraine or TTH. In a population-based study, Aaseth et al. (6) found that rhinosinusitis diagnosed without CT or nasal endoscopy was significantly associated with headache. This leads us to believe that headache sufferers in the general population may have a higher degree of sino-nasal symptoms, but may not necessarily fulfil the objective criteria for the diagnosis. A definite diagnosis of rhinosinusitis is based on a combination of sino-nasal symptoms and opacification on imaging or mucopurulent discharge or obstruction on nasal endoscopy (24). A future study using all of the necessary criteria for sino-nasal disease in the general population is needed in order to establish or refute an association between headache and sino-nasal disease.
Although evidence for an unambiguous association between headache and opacification of the paranasal sinuses is lacking, primary headache is still frequently misdiagnosed as ‘sinus headache’ (4). The patient's own desire to attribute the pain to a specific organ might be one reason for this. Lal et al. (34) showed that a large proportion of patients presenting to an otolaryngologist for sinus pressure, pain or headache had primary headache. Migraine and sino-nasal disease have overlapping symptoms: both are often unilateral and with pain in the area of the sinuses. In addition, a migraine attack is accompanied by sensitization of the peripheral neurons of the trigeminal nerves, which may give rise to pain from the maxillary division (e.g. nose and midface) of the trigeminal nerve (8). Migraine may also involve unilateral cranial autonomic symptoms such as nasal secretion during attacks (35,36), which may be interpreted as arising from inflammatory sino-nasal disease. However, the present study does not indicate that having suffered from migraine should give rise to sinus opacifications over time.
Other sino-nasal conditions such as nasal contact points may potentially cause headache (37), although they were not the aim of this study. Relevant studies on this topic often lack control groups (38) or have been performed in selected patient groups (39). Based on the lack of association between paranasal sinus opacification and headache in this study, we must underline the importance of having an adequate control group in imaging studies on sino-nasal disease, as incidental findings are indeed very prevalent.
In conclusion, migraine, TTH and unclassified headache were found not to be associated with an increased degree of paranasal sinus opacification on MRI.
Clinical implications
In a random sample of the general population, any headache, migraine, tension-type headache or unclassified headache according to the International Classification of Headache Disorders criteria was not associated with an increased degree of paranasal sinus opacification on magnetic resonance imaging. No particular paranasal sinus showed a significantly higher degree of opacification in individuals with any headache, migraine, tension-type headache or unclassified headache compared to headache-free participants.
Footnotes
Authorship contribution
AGH contributed to the conception and design, acquisition of data, analysis and interpretation of data, writing, drafting of the article and revising it critically for important intellectual content. A-SH contributed to the conception and design, analysis and interpretation of data, drafting of the article and revising it critically. WMT contributed to the conception and design, analysis and interpretation of data, drafting of the article and revising it critically. SN and VB contributed to the conception and design, analysis and interpretation of data, drafting of the article and revising it critically. LJS, KH and HBE contributed to the conception and design, acquisition of data, analysis and interpretation of data, writing and drafting the article and revising it critically.
Acknowledgements
The Nord-Trøndelag Health Study (the HUNT study) is a collaboration between the HUNT Research Centre (Faculty of Medicine at the Norwegian University of Science and Technology), the Norwegian Institute of Public Health and the Nord-Trøndelag County Council.
The present work was carried out as a collaboration between the Norwegian Institute of Public Health, the Faculty of Medicine at the Norwegian University of Science and Technology, Nord-Trøndelag County Council and the ENT Department, St. Olavs Hospital, Norway.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The authors received no financial support for the research, authorship and/or publication of this article.